PathPartner Develops Machine Learning Algorithms for Radar-Based Automotive Applications

PathPartner reduced development time from months to days. Changes to the classifier take only one hour.

Key Outcomes

  • Design changes done in days instead of months
  • Machine learning algorithms implemented and evaluated in minutes
  • Classification accuracy improved from 97% to 99.97%
  • Automotive-ready classifier completed

PathPartner provides algorithm packages for radar applications. In autonomous cars, radar-based systems can detect pedestrians and other vulnerable road users (VRUs) better than cameras at night, in inclement weather, and at greater distance.

To enable this capability, Santhana Raj, technical architect, and the Radar team developed a classifier based on radar point cloud detection. They implemented the classifier on an embedded platform and verified it in actual test scenarios.

In early testing, for example, the classifier took 5–8 seconds to detect a human—far too long to be effective. The team resolved the delay by increasing the frame time from 3 frames per second to 5 and creating a new set of features that were moving average values from the previous set of features.

Through testing and rapid design iterations, they achieved object detection accuracy of 99%. In the past, it would have taken the team 3–5 months to achieve this confidence level. With MATLAB® and Statistics and Machine Learning Toolbox™, they accomplished the project within a month.